def createMonitorChart(self, host, report_dir):
"""Create monitrored server charts."""
stats = self.monitor[host]
times = []
for stat in stats:
cycles = self.cycle_boundaries.containing_cycles(stat.time)
if cycles:
stat.cvus = max([self.cycles[cycle] for cycle in cycles])
else:
stat.cvus = 0
date = datetime.fromtimestamp(float(stat.time))
times.append(date.strftime("%H:%M:%S"))
Plugins = MonitorPlugins()
Plugins.registerPlugins()
Plugins.configure(self.getMonitorConfig(host))
charts=[]
for plugin in Plugins.MONITORS.values():
image_prefix = gnuplot_scriptpath(report_dir, '%s_%s' % (host, plugin.name))
data_prefix = gnuplot_scriptpath(report_dir, '%s_%s' % (host, plugin.name))
gplot_path = str(os.path.join(report_dir, '%s_%s.gplot' % (host, plugin.name)))
results = plugin.gnuplot(times, host, image_prefix, data_prefix, gplot_path, [640, 540], stats)
if results != None:
gnuplot(gplot_path)
charts.extend(
(name, path.replace(report_dir, '.'))
for (name, path) in results
)
return charts
def create_diff_chart(self, key, report_dir):
"""
Generate the diff chart for the specified key, comparing the
two reports for this diff.
`key`:
A section key that appears in both reports
`report_dir`:
The directory to write the data, gnuplot, and image files
"""
output_name = 'diff_{hash}'.format(hash=hashlib.md5(str(key)).hexdigest())
per_second_name = output_name + '.per_second.png'
response_times_name = output_name + '.response.png'
per_second_path = gnuplot_scriptpath(report_dir, per_second_name)
response_times_path = gnuplot_scriptpath(report_dir, response_times_name)
gplot_path = str(os.path.join(report_dir, output_name + '.gplot'))
data_path = gnuplot_scriptpath(report_dir, output_name + '.data')
labels = []
data = []
left_data = self.data1[key]
right_data = self.data2[key]
columns = set(left_data.keys()) & set(right_data.keys())
cycles = [int(v) for v in left_data['CUs']]
for column_name in columns:
labels.extend(['L_' + column_name, 'R_' + column_name])
for idx in range(len(cycles)):
values = []
for column_name in columns:
values.append(left_data[column_name][idx])
values.append(right_data[column_name][idx])
data.append(values)
with open(data_path, 'w') as data_file:
data_file.write(render_template(
'gnuplot/data.mako',
labels=labels,
data=data
))
with open(gplot_path, 'w') as gplot_file:
gplot_file.write(render_template(
'gnuplot/diff.mako',
per_second_path=per_second_path,
response_times_path=response_times_path,
use_xticlabels=not strictly_monotonic(cycles),
left_path=self.report1,
right_path=self.report2,
data_path=data_path,
key=key,
column_names=labels,
))
gnuplot(gplot_path)
return (per_second_name, response_times_name)
def createResultChart(self, key, stats, report_dir):
"""
Create a single result chart using a specified key and report directory
`key`:
The key used to identify this result. Is hashed to generate a filename
`stats`:
A map of cycle to StatsAccumulator or StatsAggregator to generate the chart from
`report_dir`:
The directory to write the data, gnuplot, and image files
Returns the relative path of the generated image in the report_dir
"""
output_name = 'results_{hash}'.format(hash=hashlib.md5(str(key)).hexdigest())
image_name = output_name + '.png'
image_path = gnuplot_scriptpath(report_dir, image_name)
gplot_path = str(os.path.join(report_dir, output_name + '.gplot'))
data_path = gnuplot_scriptpath(report_dir, output_name + '.data')
# data
labels = STATS_COLUMNS + ["E", "G", "F", "P", "U"]
data = []
has_error = False
apdex_t = self.options.apdex_t
for cycle, cycle_stats in stats.items():
values = [self.cycles[cycle]] + cycle_stats.stats_list()
if cycle_stats.errors > 0:
has_error = True
score = cycle_stats.apdex_score
apdex = ['0', '0', '0', '0', '0']
if score < 0.5:
apdex[4] = str(score)
elif score < 0.7:
apdex[3] = str(score)
elif score < 0.85:
apdex[2] = str(score)
elif score < 0.94:
apdex[1] = str(score)
else:
apdex[0] = str(score)
data.append(values + apdex)
if len(data) == 0:
# No pages finished during a cycle
return
with open(data_path, 'w') as data_file:
data_file.write(render_template('gnuplot/data.mako',
labels=labels,
data=data
))
with open(gplot_path, 'w') as gplot_file:
gplot_file.write(render_template('gnuplot/result.mako',
image_path=image_path,
chart_size=[800, 800],
maxCVUs=max(self.cycles),
datapoints=len(self.cycles),
use_xticlabels=not strictly_monotonic(self.cycles),
data_path=data_path,
has_error=has_error,
apdex_t="%0.1f" % apdex_t,
column_names=labels,
shared={}
))
gnuplot(gplot_path)
return image_name
def create_trend_chart(self, key, report_dir):
"""
Create a chart showing data trends for the specified key across all stored reports.
Returns a tuple with of 3 image paths, relative to the report dir:
(entries per second, average response time, apdex score)
`key`:
A section key that exists in self.reports_data, that specifies the data to trend
over
`report_dir`:
The path to the output directory to write the data, gnuplot, and images to
"""
output_name = 'trend_{hash}'.format(hash=hashlib.md5(str(key)).hexdigest())
per_second_name = output_name + '.per_second.png'
average_response_name = output_name + '.average.png'
apdex_name = output_name + '.apdex.png'
per_second_path = gnuplot_scriptpath(report_dir, per_second_name)
average_response_path = gnuplot_scriptpath(report_dir, average_response_name)
apdex_path = gnuplot_scriptpath(report_dir, apdex_name)
gplot_path = str(os.path.join(report_dir, output_name + '.gplot'))
data_path = gnuplot_scriptpath(report_dir, output_name + '.data')
labels = ['CUs', 'PS', 'Apdex*', 'AVG']
data = []
cycles = [int(v) for v in self.reports_data[0][key]['CUs']]
for report_idx, report_data in enumerate(self.reports_data):
for idx in range(len(cycles)):
values = [report_idx]
for column_name in labels:
values.append(report_data[key][column_name][idx])
data.append(values)
data.append(None)
with open(data_path, 'w') as data_file:
data_file.write(render_template(
'gnuplot/data.mako',
labels=labels,
data=data
))
with open(gplot_path, 'w') as gplot_file:
gplot_file.write(render_template(
'gnuplot/trend.mako',
per_second_path=per_second_path,
average_response_path=average_response_path,
apdex_path=apdex_path,
use_xticlabels=not strictly_monotonic(cycles),
data_path=data_path,
key=key,
column_names=labels,
reports_name=self.reports_name,
labels=[metadata.get('label') for metadata in self.reports_metadata],
max_cus=max(cycles),
))
gnuplot(gplot_path)
return (per_second_name, average_response_name, apdex_name)
